Skip to content
This repository has been archived by the owner on Aug 14, 2024. It is now read-only.

Latest commit

 

History

History
69 lines (60 loc) · 3.09 KB

index.md

File metadata and controls

69 lines (60 loc) · 3.09 KB
Raw
layout root permalink
lesson
.
index.html

Learn to tame your unruly data processing workflow with Snakemake, a tool for creating reproducible and scalable data analyses. Workflows are described via a human readable, Python-based language. They can be seamlessly scaled to server, cluster, grid, and cloud environments, without the need to modify the workflow definition. In this lesson, you will build up a reproducible, automated, and efficient workflow step by step with Snakemake. Along the way, you will learn the benefits of modern workflow engines and how to apply them to your own work.

The example workflow will launch several cluster jobs with the Amdahl program from the Introduction to High-Performance Computing using different numbers of processors, collect the output from each job, and create a graph of "speedup" (reference runtime, usually one processor or node or GPU, divided by the runtime with increased compute resources) as a function of the processor count. You will use this data to analyze the performance of the program, and compare it to the predictions made by Amdahl's Law. This example has been chosen over a more complex, real-world scientific workflow as the goal is to focus on building the workflow without getting distracted by the underlying science domain.

At the end of this lesson, you will:

  • Understand the benefits of workflow engines.
  • Be able to create reproducible analysis pipelines with Snakemake.
  • Estimate the proportion of parallel work from a scaling study.

Prerequisites

  • Familiarity with the command line and shell scripting, preferably having taken The Unix Shell recently or as part of this workshop.
  • A basic grasp of HPC scheduler interactions, preferably having taken Introduction to HPC recently or as part of this workshop.
  • Familiarity with the Python programming language is not required, but will help if you're curious about the inner workings of the provided "black box" programs.

Setup

Please follow the instructions in the Setup page.

The files used in this lesson can be downloaded:

  • [Linux/macOS][unix_code_pack]
  • [Windows][win_code_pack]

Once downloaded, please extract to the directory you wish to work in for all the hands-on exercises.

Solutions for most episodes can be found in the .solutions directory inside the code download.

A requirements.txt file is included in the download. This can be used to install the required Python packages. {: .prereq}

{% include links.md %}

[unix_code_pack]: {{ relative_root_path }}/files/workflow-engines-lesson.tar.gz [win_code_pack]: {{ relative_root_path }}/files/workflow-engines-lesson.zip